Steering Most Probable Escape Paths by Varying Relative Noise Intensities

We demonstrate the possibility to systematically steer the most probable escape paths (MPEPs) by adjusting relative noise intensities in dynamical systems that exhibit noise-induced escape from a metastable point via a saddle point. With the use of a geometric minimum action approach, an asymptotic theory is developed that is broadly applicable to fast-slow systems and shows the important role played by the nullcline associated with the fast variable in locating the MPEPs. A two-dimensional quadratic system is presented which permits analytical determination of both the MPEPs and associated action values. Analytical predictions agree with computed MPEPs, and both are numerically confirmed by constructing prehistory distributions directly from the underlying stochastic differential equation.